Equipment Support → Hardware By Brand → ZyXEL > USG 100 WAN SPEED

USG 100 WAN SPEED

I have the Rogers High Speed Internet service [Extreme]
Did some tests and am now questioning the integrity of the USG WAN port [no utm apps rung]

With my PC workstation connected directly to the CISCO DPC3820 Cable Modem I get 120 Mbps Down and 10 Mbps up -- this is using a JAVA applet for the speed test [no flash involved]
With my USG 100 connected to the CISCO DPC3820 my WAN Download speed drops to 51 Mbps and my Upload speed is 4 Mbps
That is quite the difference

Anyone else using Rogers Extreme with the USG 100 --- and how's your WAN/LAN/WAN specs?
--
David Mozer
IT-Expert on Call
Information Technology for Home and Business

Hmm, doesn't smell right.

Can't speak to USG100, however USG200 without UTM is delivering to the specs, see »USG200 speed tests #3

I'm assuming all tests are through wires, no wifi. Just thinking out loud, MTU issues? Physical port speed or duplex settings?

Yep I agree.

All tests are wired. My chain is
Cable modem > USG 100 >
Dell 2816 (lan)
Netgear GS105 [dmz voip]

Going to check the Dell and see all settings
--
David Mozer
IT-Expert on Call
Information Technology for Home and Business

USG 100 MUST do 100/120 while downloading and with ONLY firewall active. don't consider this speed with partial or all UTM services active, your speed will be less then 22 Mb/s!

Check the other appliances and in case of no solution, please buy all CAT6 cables, UTP if you are not in a area with high interference or STP/SSTP then connect every appliance to the "ground".

Trust me, if this is your case and you are useing cat5 and also cat5e cables and there no mistakes with configurations, then with cat6 cables you'll resolve it! I read this in many forums, valid for zyxel and for many other companies appliances! I have a usg 100 and i have cat5e cables and with only firewall I receive a file, single stream, for example a download of a exe from a browser without download managers, at 80/85 Mb/s, not more. I bought cat6 cables, and next week I'll be finally happy when they arrive home!

Interesting thesis. Cat5e cable (and in principle Cat5) in reasonable lengths with properly configured connectors should be able to do 1 Gb/s each way of pure 8b/10b encoded data transmission. However, using real messages with real protocols I have read of around 400 Mb/s being achieved. In any case, I wouldn't expect Cat5e cable to much affect USG-limited 100 Mb/s data rates, so I'll be waiting for your results, hardstyler.

kirby

Well I can confirm that its not my switches. I am using Cat5e cables. And only using the Firewall. I will try Cat6 cables and if that solves the problem I will be shocked.

[EDIT] OK, tried Cat6 cables and no difference.

I guess its time to call ZyXEL and find out why?

arg! bad news! I hope I will not have wasted money for cat6 cables.... :/ but it really solved this problem in many cases.

Anyway when mine arrive I'll post here good (hope) news!

What about duplex settings listed in either the Interface Status or through CLI as listed below:?

>configure terminal
>show port status

Does it show 100/Full or half for the WAN?

Also, try a speed test with a switch or hub in-between the cable modem and the USG

Try setting the WAN port to half duplex, and test it again.

WHY? Makes absolutely no sense to me.

My ORIGINAL question was:

and now I'm here.....bad news....cat6 cables do not solve, I get max 85 Mb/s max with firewall only or with all services active, except for idp....I really don't understand why someone solved with cat6.....I'll try the full duplex way but I need to learn the CLI so I never used it, if the change can be done by editing a configuration file I'll use this way.

We're in the process of getting clients bumped up to the new Comcast tiers (100/20 is now the same price that 50/10 used to be). We're finding that the USG100 maxes out around 48/8. The USG 200 is returning speed tests in the 115/25 range.

This is router-mode only; no IDP or AV filtering.

Edit to add: ZyXEL needs to step it up here in a massive way; we can't tell clients they need to replace a six month old $600 router because it's too slow. These things need to be built to support what comes down the pipe for 7 years or more. A faster processor is a few bucks more; no one will notice if the router is $650 in $625, but they ABSOLUTELY notice if it's obsolete in a year.

Hmmm. I would expect the USG100 to be faster than reported above.

My USG50 speed tests from Derry NH to various test servers near NY, NY generally show 15/10 on a dual WAN setup using a pair of FiOS/FAST 30/15 connections. IDP is on. I assume that the measured value is for one WAN port only, but that doesn't mean that if two ports were connected to two servers it would be any faster. I also don't assume that these speeds would necessarily be achieved using different protocols than whatever these speed test servers are using.

My speed tests are relatively consistent at various times and days of the week, but the latencies are all over the place, making some streaming video sites unwatchable at times.

kirby

An upgrade path (reduced cost) from the 100 to the 100plus would be at least expected has anyone seen any indication of this??

No worries JP, you made much sense to me. Throw in the fact that managers (with bean counters) not engineers probably make the decisions.

In my own personal experiences, I've seen a range of 60-80 Mbps down on the USG 100. The USG50 performs slightly better in terms of throughput but that is most likely due to it having a faster processor. I am starting to wonder if ZyXEL didn't think internet speeds were going to increase the way they have over the last few years. What was acceptable throughput back in 2007 seems to no longer fit the needs of customers as ISP's are able to offer more speeds than the lower end USG's can handle.

Last Night I put the EdgeRouter Lite [CDN $140 -- includes everything] into service replacing my USG100 for comparisons sake and I can report that I'm getting 100Mbps Down and 10Mbps Up WAN/VLAN/WAN on my wired workstation sitting in a Dell PowerConnect 2816 switch with VLAN's configured.

Is it a fair comparison? Not sure it is cause I'm not intimately familiar with how my EdgeRouter Lite Firewall is configured and all the interactions. I'm using a config I picked up from the EdgeMAX Forum so I need to learn how to properly configure this unit and make sure all the protection I need is actually in place. I may have to bounce off Brano since he is the Linux GURU and also has a good opinion of the EdgeRouter.
--
David Mozer
IT-Expert on Call
Information Technology for Home and Business

Well JP, there is the responsibility of the purchaser to ensure forward room for growth but also on the vendor to deliver specifications. If they failed on the latter then they should provide an upgrade path.
Doesn't the usg100 state 225? Which should easily handle any combination up to 225 and I will give up 25 for overhead so make it 200. Thus 100 up and 100 down should be supported or a 125 75 combo etc............

So why come up with a usg100+
a. they recognized ISP speeds have outpaced their products.
b. perhaps they know that the 100 is underperforming.

The problem with the specification is that the vendor can prove that they do 225 or 100 or 1000 what ever. Its not like they did not measure it, its just what they measure and what you measure.

For instance 100Mbps Ethernet is not 100Mbps IP traffic, its about 94 Mbps. So when a vendor says throughput is 100Mbps and your test says 94Mbps. The spec is right and your are testing wrong in regards to the spec. And there is difference between UDP and TCP and zzzz

And are we talking about a single session or a multiple concurrent sessions?

Could they clarify, sure, will they, probably not, unless somebody with superior performance does it.
Its not hard to list

L2 performance : Wirespeed ie 100Mbps single direction. 200Mbps bi-directional.
L3 performance : 94 Mbps or 50 or 12 or 99 etc.
L4 performance : 67 or 13 or 88
--
"Perl is executable line noise, Python is executable pseudo-code."

Yes but no one is complaining about usg 50 spec vs actual or USG 200 spec vs actual. Thus am concluding there is something different and off about the USG100.

i'm with you all: companies must provide detailed performance in the datasheets, of course they are only lab tests but they will be more more then a simple test by rfc 2544 udp 1500 ecc.... values. but all never do that cause all products must be destroyed, ehm...never built cause they are not in line with the actual ISP speeds. 2 months ago I downloaded only 4 torrent at the same time maxing my isp speed, only 12 Mb/s in downloas so is a ridicolous speed...then...my usg 100 with all utm services activated showed cpu at 95% constantly!!! it cannot provide fast log alerts and also it cannot make and send the daily report cause is too busy.

usg 100 plus is great but is a 100 equal in utm throughput...

edge router lite? great product but it is a simple router and fast is obvious....utm features? none. so it cannot be compared with usg 100 series of zyxel... my netgear dgn2200 is a modem adsl2plus that is also a firewall and a router but has a ridicolous firewall, limited rules and ultra low protection and basic to nothing about logs...but it can do 100 Mb/s, 15 more of the usg 100....but it is a product that we cannot compare with a utm firewall.

when will be released the 3.30 fw for the usg? I hope in a better IDP supporting recent OS, they said yes but today no news about latest OS support...

It's perhaps worth noting that we've out 50-75 ZyXEL routers into the field over the last 15 years. Probably 25 are still in service. Of those no longer in service, ONE failed. Every other one was retired because it was too slow for an upgraded Internet connection.

But telling a client they need to replace a $200-$300 ZyWALL 10 or 10+ that's been in service for six years is a different conversation that having to replace a $600 USG 100 that's been in service less than a year because it can't keep up with their $100/month 50/10 Internet connection. What's happening is that the ZyXEL's model of pricing routers based on throughput isn't working now that Comcast and other ISPs are offering blistering fast (100/20 for $190/month) connections for business customers. Maybe they need to look at a different tiering/pricing model.